AI-first physical cooking infrastructure
The Physical Execution Layer for AI Cooking
On2Cook turns culinary intent into machine-executable cooking curves, then controls the heat, moisture, motion and time needed to produce the outcome.
01 / AI execution system
How On2Cook senses, decides and cooks
The intelligence is valuable because it has a protected physical actuator. Select a stage to see how intent becomes an actual cooking action.
Future image: the requested dish beside the ingredient load and target portion.
Dish intent
A recipe request begins as culinary intent: the dish, ingredient state, quantity, cuisine, texture and operator constraints.
- Component and vessel temperatures
- Current, voltage and power behavior
- Altitude, ambient and site profile
- Portion, load and ingredient state
- Microwave and induction power
- Stirring pattern and motion
- Water and fluid dosing
- Timing, pause, derating and stop
02 / Intelligence maturity
From programmed recipes to autonomous cooking
One architecture, presented with the status discipline investors need.
Deployed now
Capabilities already used to execute standardized recipes in commercial kitchens.
- Structured recipe packages with step-by-step machine actions
- Coordinated microwave, induction, stirring, water and timing
- App-delivered recipes with deterministic device execution
- 600+ standardized recipes across an installed commercial base
Recipe source, flavor model, ingredient model, process model and execution graph.
Commercially validated through 380+ device sales and 80+ brands.
03 / Recipe intelligence evidence
One recipe. Multiple electrical conditions. Visible outcomes.
A 100 g Moong Dal Halwa study records temperature and electrical traces alongside the cooked result. It is an early engineering evidence set for why recipes need context, sensing and edge control, not only a static timer.
A direct-line baseline for comparison
T2 records the 100 g recipe on a direct 234 V line. The later comparison slides use this run as the reference trace.
Source: “Moong Dal Ka Halwa on different voltages graphs and outputs_sandyv1”, June 2026. These are engineering trial records, not a claim that full autonomous compensation is commercially released.
04 / Patented actuator
Heat + moisture + motion + time
On2Cook combines deep volumetric microwave energy with surface heat, autonomous stirring and fluid injection. That turns cooking physics into a software-addressable system.
05 / Defensibility
A layered patent fence around the complete system
The portfolio follows the engineering sequence of the product. Select a layer to see the problem addressed, the filed technical approach and the route it closes around the platform.
Heat-source orchestration
Pan heat creates browning, frying and flavor development, while microwave energy heats through the food mass. Operating them as unrelated heaters does not create a controlled hybrid cook.
The filed architectures apply traditional heat or induction and microwave energy to the same food, with open and closed states, sensors, stirring, moisture and mode control coordinated around the recipe.
Establishes the inside-and-outside heating principle and the practical microwave-plus-induction architecture at the centre of On2Cook.
Route this layer closesCopying two heat sources without the coordinated vessel, modes and process control.
- Portfolio scope
- Blocks 01-02 / 2010 foundation and related multi-source family
- Representative identifiers
- IN3051/DEL/2010; WO2012085944A1; US9402511B2; GB2498916B; US10022021B2; WO2018025095A1
Source: On2Cook Patent Portfolio Partner Brief, August 2026. Identifiers are representative, not an exhaustive national-phase docket. Status and legal scope depend on the issued or pending claims and the official records in each jurisdiction; this summary is not a legal or freedom-to-operate opinion.
06 / Compounding intelligence
Every cook expands the intelligence
The recipe graph is the bridge between culinary know-how, device control and fleet learning.
07 / Kitchen evidence
Cooking evidence, not cooking claims
Real dishes across cuisine families, meal occasions and cooking formats, paired with the operating outcomes presented in the technology deck.
Selected ranges are presented in the December technology and cooking-advantages deck; validation basis varies by dish, load, kitchen and comparison method.
07B / AI-assisted operations
AI-assisted cooking changes the kitchen operating model
The AI recipe and control layer is the mechanism. Lower manual intervention, concurrent cooking and the consolidation of equipment and footprint are the kitchen-level effects.
Operating scenarios shown on slides 2, 4 and 6 of the December technology and cooking deck. Actual staffing, throughput and cost vary by menu, batch size, tariffs and site conditions.
08 / Commercial architecture
How the installed base generates revenue
Hardware opens the account. Select each commercial lane to see what compounds and the deck evidence already behind it.
B2B device + enterprise
Customer categories, deployed brands and repeat-sale routes show where the current hardware engine is already working.
- Who pays
- QSRs, cloud kitchens, hotels, cafes, caterers and institutions
- What they buy
- Devices, installation, rollout support, accessories and service
- Revenue shape
- Upfront hardware plus installed-base recurring revenue
- Evidence / trigger
- ₹2.048 Cr B2B direct lifetime revenue and ₹1.000 Cr enterprise lifetime revenue in the latest deck.
Hardware margin funds the intelligence layer
The device creates the installed base; software, recipes, service and data can then expand lifetime value without requiring a new hardware sale for every improvement.
- India selling price
- ₹2,00,000 + GST
- Manufacturing cost
- ₹58,000 + GST
- Gross margin
- 60-70%
- Global selling price
- US$3,500
09 / Strategic validation
Antunes validates the global partnership model
A 70+ year US equipment company provides external validation, co-development capital and a route into global food-service networks.
- Non-dilutive co-development relationship
- Adjacent products including steamers, fryers, toasters and larger units
- Royalty income forecast to begin Q4 FY2027
- Certification, service and global-channel acceleration
10 / Commercial proof
Traction with a clean line between revenue and pipeline
Recognized, contracted and projected values are deliberately separated.
Brands, certification and revenue belong in the same proof frame
The latest deck connects 380+ device sales and 600+ standardized recipes with named food-brand adoption, CE certification and the revenue visibility shown above.
₹1,563.71 Cr five-year revenue build
The model combines core B2B hardware, Antunes-linked products and services, AMC, accessories and later partner royalty streams.
Forward-looking management projections from the August 2026 investor deck. They are not guidance or recognized revenue.
11 / Pre-Series A
What the ₹20 Cr round accelerates
The current CCPS round closes the gap between proven invention and scaled AI-first infrastructure.
Deployment of the remaining ₹10.38 Cr
On2Cook investment opportunity